by Calculated Risk on 1/08/2016 08:01:00 PM

A few excerpts from a research note by Merrill Lynch economists Michelle Meyer and Alexander Lin: Melting snowmen

After two brutal winters, we have been enjoying an unprecedented warm winter start. The average temperature in December was 38.6 degrees, versus the previous record of 37.7 and the historical average of 33.2. If we create a national aggregate for temperature which is weighted by state population instead of area, we see an even bigger divergence from the norm in December. The appeal of using population weights is that it will put more emphasis on the temperature in the areas which have a greater economic contribution.

Another proxy for gauging the weather in the winter is to show heating degree days, which measures the demand for energy to heat houses or businesses ... the deviation from the norm for heating degree days in December and this past month was literally off the charts.
...
In a research note released February last year, Chicago Fed economists Justin Bloesch and François Gouri provide an insightful methodology for examining the impact of weather on economic indicators. Using daily data from the more than 1,200 weather stations that make up the US Historical Climatology Network, they constructed temperature and snowfall indices based on deviations from historical levels. ... They test the impact of this deviation on economic indicators – both in winter months and in April, May and June, to understand the potential payback. There is strong evidence of a weather impact on nonfarm payrolls as well as claims, housing starts, and permits. Focusing on payrolls, they estimated that a 1 standard deviation increase in temperature resulted in a 0.04% boost to nonfarm payroll growth during the current month, which would translate into roughly 60,000 additional jobs at today’s levels. Similarly, a 1 standard deviation decline in snowfall would result in a 0.03% bump to job growth, or 45,000 additional jobs.

Plugging in the December temperature and snowfall data to the output of the model from Bloesch and Gouri, we find that the weather can explain about 97,000 of job growth. This would imply that without the weather distortion, the economy would have added 195,000 jobs. While we don’t want to give a false sense of precision, this seems like a reasonable approximation. Before the sharp acceleration the past three months, job growth was trending at around 200,000 a month.
emphasis added